Fix dataset_info.features: declare script_type column

README.md

@@ -31,6 +31,8 @@ dataset_info:
             '5': Uchen
             '6': multiscript
             '7': non_tibetan
+    - name: script_type
+      dtype: string
   splits:
     - name: test
       num_bytes: 0
@@ -69,6 +71,7 @@ All images are BDRC manuscript page scans, balanced by subclass.
 | `id` | string | BDRC page id (e.g. `W00KG09391-I00KG093950005`) |
 | `image_bytes` | binary | JPEG/PNG page image |
 | `script` | string | One of: `Danyig`, `Druma`, `Gyuyig`, `Pedri`, `Tsugdri`, `Uchen`, `multiscript`, `non_tibetan` |
+| `script_type` | string | Sub-script / subclass name (e.g. `DraDring`, `Multi-Scripts`, `Other`) |
 
 ## Load in Python